Skip to content

fix: remove unused fp8 training args in config #1018

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
terrykong merged 1 commit into from
Aug 29, 2025
Merged

fix: remove unused fp8 training args in config #1018

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 8 additions & 7 deletions examples/configs/sft_openmathinstruct2_megatron.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,13 +88,14 @@ policy:
env_vars:
PYTORCH_CUDA_ALLOC_CONF: "expandable_segments:False"

fp8_cfg:
enabled: true
fp8: hybrid
fp8_recipe: delayed
fp8_param: true # false gives the following error: "RuntimeError: /TransformerEngine/transformer_engine/common/gemm/cublaslt_gemm.cu:116 in function CanonicalizeGemmInput: Assertion failed: !is_fp8_dtype(ret.Atype). Input A is missing column-wise usage"
fp8_dot_product_attention: false #true
fp8_multi_head_attention: false #true
## fp8 training currently not supported
#fp8_cfg:
# enabled: true
# fp8: hybrid
# fp8_recipe: delayed
# fp8_param: true # false gives the following error: "RuntimeError: /TransformerEngine/transformer_engine/common/gemm/cublaslt_gemm.cu:116 in function CanonicalizeGemmInput: Assertion failed: !is_fp8_dtype(ret.Atype). Input A is missing column-wise usage"
# fp8_dot_product_attention: false #true
# fp8_multi_head_attention: false #true

dynamic_batching:
enabled: false
Expand Down